A Data-Driven Adhesive Adhesion Parameter Adaptive Adjustment Method and System

Abstract

This invention discloses a data-driven adaptive adjustment method and system for adhesive adhesion parameters, belonging to the field of adaptive adjustment technology for adhesive adhesion. It includes constructing a stroke-pressure feasible region based on hardware constraints and order requirements, and generating multiple sets of adhesive adhesion adjustment parameters by combining perturbation expansion. During vacuum partition bonding, a collaborative control mechanism of motor torque and pressure head termination position is introduced to achieve multi-condition coverage and dynamic execution of the bonding process. Furthermore, a state-reward cumulative scoring model is constructed by collecting displacement, torque, and contact state data, and a three-dimensional evaluation surface is established based on the maximum downward stroke and unit pressure for extreme value search, thereby achieving automatic selection and application of optimal bonding parameters. This solves the problems of reliance on geometric empirical rules, poor parameter transferability, and unstable adhesion quality in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of adaptive adjustment technology for adhesive adhesion, and more specifically, to a data-driven adaptive adjustment method and system for adhesive adhesion parameters. Background Technology

[0002] In electronic module bonding processes, adhesive adhesion adjustment technology is mainly used for interface bonding control of flexible or semi-flexible devices such as FPCs and display modules. Existing technologies typically combine zoned pressure control with vacuum-assisted bonding. By dividing the bonding area into multiple independent control units and using displacement sensors to obtain substrate surface warping information, differentiated loading of the pressure head is achieved for each area, thereby improving the problem of uneven local bonding. Simultaneously, pre-evacuation of the vacuum chamber reduces residual gas at the interface, and the pressure head array performs staged pressing to control the bonding process from initial contact establishment to subsequent pressure holding and curing. However, this type of method essentially still relies on a geometry-driven parameter mapping mechanism and lacks the ability to directly model the viscoelastic changes of the adhesive material during the pressure process.

[0003] In the aforementioned bonding process, maximum downward stroke and unit pressure are two core control parameters. Maximum downward stroke constrains the physical advance range of the indenter in the vertical direction, determining whether the indenter can overcome the substrate warpage and achieve global contact coverage. Unit pressure describes the average normal load intensity applied by the indenter to the contact interface, directly affecting the wetting and spreading ability of the adhesive layer and the tightness of the interface adhesion. Together, these two parameters determine the balance between the indenter's "contact establishment ability" and "adhesive layer flowability," and are key process variables affecting the final adhesion effect.

[0004] Both the maximum downward stroke and the unit pressure have a significant impact on the adhesive adhesion conditioning process. On the one hand, while a larger downward stroke is beneficial for improving contact coverage, it can easily introduce the risk of over-compression. On the other hand, while a higher unit pressure helps improve initial wetting and eliminate microvoids, it may lead to adhesive flow instability and stress concentration. Therefore, both parameters exhibit bidirectional constraints; their values ​​simultaneously affect contact integrity, adhesive layer stability, and residual stress distribution, thus creating a coupled effect on the final adhesion quality.

[0005] When the maximum downward stroke is set too low, the pressure head cannot fully overcome the local warpage height of the substrate, resulting in preferential contact in higher areas while lower areas remain in a gap state. This easily leads to the formation of local suspended structures at the interface, causing problems such as residual air bubbles and insufficient effective bonding area. At the same time, due to insufficient contact establishment, the subsequent adhesive layer expansion path is limited, resulting in a discontinuous distribution of the overall adhesion interface and reducing bonding consistency.

[0006] When the maximum downward stroke is set too high, the indenter will apply excessive displacement compression to the substrate, causing the adhesive layer to be over-compressed in the vertical direction. This results in uncontrolled overflow of the adhesive towards the edge areas, and the flexible substrate may experience irreversible deformation or localized stress concentration. In addition, excessive stroke can cause the indenter to enter an "overstroke compression" state, keeping the interface under high pressure for a long time, increasing the risk of subsequent edge warping and structural fatigue.

[0007] When the unit pressure is low, the normal driving force of the indenter on the adhesive layer is insufficient, causing the adhesive to fail to fully wet the microscopic surface structure. The process of interfacial contact expanding from point contact to surface contact is hindered, easily leading to virtual contact or weak adhesion. At the same time, due to insufficient driving force, the gas inside the adhesive layer is difficult to expel effectively, further increasing the probability of air bubble residue and thus reducing the overall adhesion strength.

[0008] While higher unit pressure can enhance interfacial wetting and rapidly eliminate micropores, it can also cause excessive lateral flow of the colloid, resulting in uneven colloid thickness distribution and adhesive overflow at the edges. Furthermore, high pressure introduces significant residual stress within the colloid layer, making the interface more susceptible to edge lifting or peeling failure under subsequent temperature changes or mechanical loads.

[0009] In practical control processes, existing technologies typically only set the maximum downward stroke and unit pressure individually based on the substrate's geometric warp or empirical rules, lacking a systematic optimization and global search mechanism for the coupling relationship between the two, and failing to establish a continuous evaluation model based on experimental feedback. Therefore, when different batches of adhesives, different environmental conditions, or different structural dimensions change, the parameter adaptability is weak, easily leading to large fluctuations in bonding quality, manifesting as problems such as residual bubbles, edge lifting, and unstable peel strength.

[0010] To address the above problems, this invention proposes a solution. Summary of the Invention

[0011] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a data-driven adaptive adjustment method and system for adhesive adhesion parameters. By constructing a stroke-pressure feasible region based on the maximum downward stroke and unit pressure, and introducing a bidirectional disturbance and contact-pressure-torque collaborative control mechanism, and combining state-reward cumulative evaluation and continuous evaluation surface extreme value search, the adaptive optimization of adhesive adhesion parameters from empirical settings to data-driven optimal solutions is achieved, thereby solving the problems of reliance on geometric empirical rules, poor parameter mobility, and unstable adhesion quality in the prior art.

[0012] To achieve the above objectives, the present invention provides the following technical solution: A data-driven adaptive adjustment method for adhesive adhesion parameters includes the following steps: acquiring hardware constraint data and order demand data, constructing a stroke-pressure feasible region, and generating a combination pair of basic maximum downward stroke and basic unit pressure through constraint optimization to construct a basic adhesive adhesion adjustment group; performing bidirectional continuous perturbation on the basic adhesive adhesion adjustment group in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacies and adhesive flow states; dividing the maximum downward stroke into a contact establishment stage and a compaction stage based on the gradual expansion of the contact area during the pressure head's downward movement; adjusting the motor torque output during the contact establishment stage and maintaining the motor torque during the compaction stage. The output is stabilized, thus converting the adhesive adhesion adjustment group into motor torque and pressure head termination position; each second adhesive adhesion adjustment group is converted and loaded into the partition pressure control unit respectively, and electronic module bonding is performed under vacuum-assisted bonding conditions; features are extracted from the bonding process data, and a reward signal is constructed based on the bonding results. Adhesive adhesion score corresponding to each second adhesive adhesion adjustment group is generated through state-reward cumulative evaluation; the maximum downward stroke and unit pressure in each second adhesive adhesion adjustment group are used as two-dimensional inputs, and the corresponding adhesive adhesion score is used as the output to fit and construct a three-dimensional evaluation surface. The optimal adhesive adhesion adjustment group is screened through surface extremum search and applied to the electronic module bonding process.

[0013] A data-driven adaptive adjustment system for adhesive adhesion parameters includes an adhesive adhesion initialization module, an adjustment disturbance generation module, a torque mapping module, a partitioned execution module, a feedback scoring module, and a surface optimization module. The adhesive adhesion initialization module acquires hardware constraint data and order demand data, constructs a stroke-pressure feasible region, and generates a combination of the basic maximum downward stroke and the basic unit pressure through constraint optimization to construct a basic adhesive adhesion adjustment group. The adjustment disturbance generation module continuously perturbs the basic adhesive adhesion adjustment group bidirectionally in both the stroke and pressure dimensions, generating several second adhesive adhesion adjustment groups covering different contact adequacies and adhesive flow states. The torque mapping module divides the maximum downward stroke into a contact establishment stage and a compaction stage based on the gradual expansion of the contact area during the pressure head's downward movement. The system establishes a stage to adjust the motor torque output, maintaining stable motor torque output during the compaction stage, thereby converting the adhesive adhesion adjustment group into motor torque and pressure head termination position. A zone execution module converts each second adhesive adhesion adjustment group and loads it into a zoned pressure control unit, executing electronic module bonding under vacuum-assisted bonding conditions. A feedback scoring module extracts features from bonding process data and constructs a feedback signal based on the bonding results, generating an adhesive adhesion score for each second adhesive adhesion adjustment group through state-feedback cumulative evaluation. A surface optimization module uses the maximum downward stroke and unit pressure of each second adhesive adhesion adjustment group as two-dimensional inputs, and the corresponding adhesive adhesion score as the output to fit and construct a three-dimensional evaluation surface. It then uses surface extremum search to select the optimal adhesive adhesion adjustment group for application in the electronic module bonding process.

[0014] The present invention also includes an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform a data-driven adaptive adjustment method for adhesive backing parameters.

[0015] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements a data-driven adaptive adjustment method for adhesive backing parameters.

[0016] The technical effects and advantages of the data-driven adaptive adjustment method and system for adhesive backing parameters of this invention are as follows: 1. This invention constructs a stroke-pressure feasible domain and combines hardware constraints and order requirements to jointly screen the maximum downward stroke and unit pressure, making the initial parameter set physically feasible and process-adaptable. At the same time, it expands the parameter distribution range through bidirectional continuous perturbation, expanding the adjustment group from a single empirical point to a multi-dimensional parameter set covering the differences in contact adequacy and adhesive flow state. This significantly improves the expressive power and coverage density of the parameter space, so that subsequent optimization no longer relies on a single empirical setting, but is based on a systematic search of the complete feasible domain, improving the adaptability and stability under different substrate warpage conditions and different order specifications.

[0017] 2. This invention achieves segmented control of the contact establishment and compaction stages, and introduces a combined control mechanism of motor torque and pressure head termination position. This allows the pressure head to gradually establish force during contact expansion and maintain a stable compaction state during the compaction stage, thereby reducing abrupt stress changes in the adhesive layer. Simultaneously, through a state-reward cumulative evaluation mechanism based on the bonding process state chain and quality feedback signals, multi-dimensional quality results such as bubbles, edge lifting, and peel strength are transformed into a unified scoring scalar. Furthermore, a three-dimensional evaluation surface of stroke and pressure is constructed. Stable optimal adjustment parameters are screened through extreme value search and neighborhood consistency verification, transforming parameter selection from offline empirical adjustment to a data-driven global optimal search process. This effectively reduces problems such as bubble residue, edge lifting, and uneven adhesive layer stress, improving bonding consistency and process robustness. Attached Figure Description

[0018] Figure 1 This is a schematic flowchart of a data-driven adaptive adjustment method for adhesive backing parameters according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1, Figure 1 This invention presents a data-driven adaptive adjustment method for adhesive adhesion parameters, comprising the following steps: S1. Obtain hardware limitation data and order demand data, construct the stroke-pressure feasible region, and generate a combination pair of basic maximum downward stroke and basic unit pressure through constraint optimization to construct the basic adhesive adhesion adjustment group.

[0021] In this embodiment, the process of acquiring hardware constraint data and order demand data, constructing a stroke-pressure feasible region, and generating a combination pair of basic maximum downward stroke and basic unit pressure through constraint optimization to construct a basic adhesive adhesion adjustment group specifically involves: The hardware limitation data includes the upper limit of the stroke of the pressure head structure of the electronic module bonding equipment and the rated torque of the drive motor; the order requirement data includes the substrate thickness and warpage statistical range in the order. The maximum allowable pressing depth set is determined based on the upper limit of the pressure head structure stroke, and the maximum output contact pressure set is determined based on the rated torque of the drive motor. The maximum allowable pressing depth set and the maximum output contact pressure set are then combined using a Cartesian combination to generate the initial stroke-pressure combination space. The initial stroke-pressure combination space is screened based on the substrate thickness and warpage statistical range. Stroke combinations that are less than the maximum warpage height of the substrate and correspond to insufficient contact are eliminated, and pressure combinations that exceed the allowable range of substrate compression deformation and correspond to overpressure risk are eliminated, thus obtaining a set of feasible combinations that meet the physical contact conditions. In the feasible combination set, the unit pressure and the contact coverage of the pressure head are used as constraints. The contact coverage is set to be no less than the effective bonding ratio of the substrate surface. Combinations with unit pressure exceeding the critical range of substrate compression are eliminated, while the combination range that meets the coverage and compression limits is retained to form the stroke-pressure feasible region. Within the feasible range of stroke-pressure, the difference between the pressure head termination displacement and the remaining thickness of the substrate after compression is used as the screening criterion. Combination points whose difference falls within the preset allowable range are preferentially retained. The combination points are then discretely selected to generate combination pairs of the maximum downward stroke and the unit pressure of the base, thus constructing the base adhesive adhesion adjustment group.

[0022] In this embodiment, the hardware constraint data is physical boundary information used to characterize the electronic module bonding device in terms of structure and driving capability. The upper limit of the pressure head structure stroke indicates the maximum vertical downward displacement range that the pressure head can achieve under the constraints of the mechanical guide rail and the limiting structure. The rated torque of the drive motor indicates the maximum driving torque capability that the motor can stably output under continuous operation. These two types of data are used together to limit the physical space boundary that the bonding process can execute.

[0023] In this embodiment, the order demand data is the physical state constraint information of the substrate obtained from the statistics of different product orders. The substrate thickness represents the nominal structural thickness range of the substrate to be bonded, and the warpage statistical range represents the distribution range of surface height deviation formed by the substrate during processing and transportation. This data is used to describe the non-ideal flatness of the substrate before bonding, and thus serves as the basis for the adaptation of stroke and pressure settings.

[0024] In this embodiment, the maximum allowable pressing depth set is the range of executable pressing displacements calculated based on the upper limit of the indenter structure stroke. This set is obtained by breaking down the mechanical limits of the device into multiple executable discrete displacement levels, and is used to limit the space of selectable pressing depths of the indenter under the condition that no mechanical interference or structural damage occurs.

[0025] In this embodiment, the maximum output contact pressure set is the stable output pressure range obtained by converting the rated torque of the drive motor. This set is obtained by mapping the motor output capacity to the equivalent pressure level formed by the pressure head contact interface load, and is used to limit the pressure selection space that can be applied to the substrate surface under the condition of not exceeding the motor load capacity.

[0026] In this embodiment, the initial stroke-pressure combination space is a complete combination set formed by pairing the maximum permissible pressing depth set and the maximum output contact pressure set one by one. This combination space represents all possible stroke and pressure combinations under the extreme conditions of the equipment, and is used for further limitation in combination with substrate constraints.

[0027] In this embodiment, the stroke combination that corresponds to insufficient contact at the maximum warpage height of the substrate is eliminated as the most unfavorable contact condition determined based on the maximum height in the warpage statistical range. When the pressure depth of the indenter is less than the warpage height, it will not be able to fully contact the low-lying area of ​​the substrate. Therefore, this type of stroke combination will physically lead to local suspension or false contact, and thus it will be removed.

[0028] In this embodiment, the pressure combination that corresponds to the risk of overpressure in the allowable range of substrate compression deformation is eliminated by the maximum unit pressure limit determined based on the compression characteristics of the substrate material. When the unit pressure exceeds this range, it will cause irreversible deformation or structural damage to the substrate. Therefore, this type of pressure combination is determined to be unexecutable in the actual bonding process.

[0029] In this embodiment, the pressure head contact coverage rate is a physical quantity that represents the proportion of the effective bonding area actually formed on the substrate surface under the action of the pressure head. This index is used to measure whether the pressure head achieves uniform contact across the entire area, where a higher coverage rate indicates a more complete effective bonding area.

[0030] In this embodiment, the effective bonding ratio requirement of the substrate surface is the minimum contact coverage requirement set according to the order quality standard. This requirement is used to ensure that there are no obvious gaps or local detachment areas in the product after bonding, thereby constraining the combination of stroke and pressure to meet the overall coverage conditions.

[0031] In this embodiment, the combination restriction where the unit pressure exceeds the critical range of substrate compression is a rule that constrains the degree of local pressure concentration based on the coverage condition. When the unit pressure forms a concentrated load in a local area that exceeds the material's withstand threshold, the combination will be restricted from entering the subsequent feasible region even if the coverage condition is met.

[0032] In this embodiment, the stroke-pressure feasible region is a stroke and pressure combination subspace formed after satisfying the mechanical limits of the equipment and the physical bearing capacity of the substrate, as well as the contact coverage requirements and compression constraints. This subspace represents the range of all safe parameter combinations that can be actually used for the bonding process.

[0033] In this embodiment, the difference between the pressure head termination displacement and the remaining thickness of the substrate after compression is a quantitative indicator used to describe the degree of matching between the final pressing position of the pressure head and the actual residual structural thickness of the substrate after compression. This indicator is used to determine whether the pressure head completes the bonding within a reasonable range without producing over-pressure or under-pressure states.

[0034] In this embodiment, the preset allowable range is a reasonable matching range between the pressure head termination position and the residual thickness of the substrate obtained from the statistical analysis of historical bonding good product data. This is used to constrain the final selected stroke and pressure combination to be within the process range corresponding to the stable bonding state.

[0035] It should be noted that this embodiment unifies the physical boundary constraints of the equipment with the state constraints of the substrate, so that the stroke and pressure are no longer set as independent empirical parameters, but converge to the executable parameter space through multi-layer constraints, thereby avoiding the problems of overpressure damage or insufficient contact caused by relying solely on empirical settings in traditional bonding processes.

[0036] It should be noted that this embodiment constructs a stroke-pressure feasible domain and introduces dual constraints of substrate warpage and compression, so that the parameter selection process has a physical correspondence consistent with the actual bonding results, thereby improving the stability and consistency of parameter adaptation under different order conditions.

[0037] It should be noted that the beneficial effect of this embodiment is that by uniformly mapping hardware limits, substrate conditions and contact quality requirements into an executable stroke-pressure combination space, the subsequent optimization process can select parameters within a safe boundary, thereby significantly reducing the probability of bonding defects and improving the controllability and consistency of the electronic module bonding process.

[0038] S2, for the basic adhesive adhesion adjustment group, bidirectional continuous perturbation is performed in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacy and adhesive flow states.

[0039] In this embodiment, the basic adhesive adhesion adjustment group is subjected to bidirectional continuous perturbation in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacies and adhesive flow states, specifically as follows: The bidirectional continuous disturbance extends the displacement along the allowable range of the mechanical clearance of the pressure head, and the pressure extends the load along the range of the output capacity of the drive motor. For the maximum downward stroke in the basic adhesive adhesion adjustment group, the displacement is extended in the increasing and decreasing directions according to the preset step size along the allowable range of the mechanical clearance of the pressure head to obtain the stroke disturbance sequence; For the unit pressure in the basic adhesive adhesion adjustment group, the load is expanded in the increasing and decreasing directions according to the preset step size along the output capacity range of the drive motor to obtain the pressure disturbance sequence. The stroke disturbance sequence and the pressure disturbance sequence are paired one by one to form a stroke-pressure disturbance combination set. The contact compensation adjustment is performed on the indenter termination displacement corresponding to each combination so that the termination displacement is not less than the maximum warping height of the substrate and does not exceed the upper limit of the indenter structure stroke. At the same time, the unit pressure corresponding to each combination is adjusted to ensure that the unit pressure does not exceed the pressure limit corresponding to the compressibility range of the substrate. After completing the displacement compensation adjustment and pressure loading limit adjustment, each combination is discretely selected at preset intervals to obtain several combination points covering different pressing depths and loading pressure ranges. Then, using the maximum pressing stroke and unit pressure corresponding to each combination point as parameter pairs, several second adhesive adhesion adjustment groups are constructed.

[0040] In this embodiment, the basic adhesive adhesion adjustment group is a set of basic control parameters obtained from the constraint optimization in the previous stage. This parameter pair consists of the maximum pressing stroke and the unit pressure. The maximum pressing stroke is the deepest pressing position that the indenter is allowed to reach without structural interference. The unit pressure is the equivalent loading intensity applied by the indenter to the unit contact area. This set of parameters is used as the reference starting point for the generation of subsequent disturbances.

[0041] In this embodiment, the bidirectional continuous disturbance is a process of continuously expanding the parameters along the increasing and decreasing directions, with the maximum downward stroke and unit pressure in the basic adhesive adhesion adjustment group as the center. The increasing direction represents gradually increasing the pressure depth or loading intensity of the pressure head, and the decreasing direction represents gradually decreasing the corresponding parameters. The parameter distribution that covers the boundary changes of the covering process is formed through the bidirectional expansion, which is used to enhance the coverage of different bonding states.

[0042] In this embodiment, the allowable range of mechanical clearance of the pressure head is the maximum displacement range that the pressure head can achieve under the joint constraints of the mechanical guide rail, frame and limiting structure. This range is determined by the upper limit of the equipment structure and the safety limit switch. Its function is to limit the maximum downward stroke to expand within the condition that no mechanical collision or structural jamming will occur.

[0043] In this embodiment, the output capability range of the drive motor is the equivalent load capability range corresponding to the stable torque output that the motor can provide under continuous operation conditions. This range is jointly limited by the motor's rated current, heat dissipation capacity, and driver limitations, and is used to constrain the expansion of unit pressure without causing motor overload or thermal instability.

[0044] In this embodiment, the stroke disturbance sequence is a set of discretized displacement parameters obtained by gradually expanding to the upper and lower limits with a fixed increment centered on the basic maximum downward stroke. This set is used to express the possible control points of the indenter contact state change under different indentation depths.

[0045] In this embodiment, the pressure disturbance sequence is a set of discretized loading intensities that expands in fixed increments towards high and low pressure directions within the driving capability range, with the basic unit pressure as the center. This set is used to express the control points of the stress state change of the adhesive layer under different compression intensities.

[0046] In this embodiment, the stroke-pressure disturbance combination set is a set of parameter pairs formed by pairing each displacement value in the stroke disturbance sequence with each pressure value in the pressure disturbance sequence in a one-to-one manner. Each combination pair represents a complete pressure head pressing and loading condition, which is used to cover different contact states.

[0047] In this embodiment, the contact compensation adjustment is to address the situation where the maximum downward stroke in the combination is insufficient to cover the highest warped area of ​​the substrate. The stroke parameter is increased and corrected so that the actual termination position of the pressure head can ensure that it first contacts the highest point area of ​​the substrate, thereby avoiding a local suspended and uncontacted state.

[0048] In this embodiment, the loading limit is adjusted to constrain and correct the pressure parameters when the unit pressure in the combination exceeds the compressibility of the substrate, so that the loading intensity is limited to a safe range that will not cause irreversible deformation of the substrate or excessive extrusion of the adhesive layer.

[0049] In this embodiment, the preset quantity interval is a method of selecting parameter points in the adjusted combination set according to a fixed step size or a fixed ratio interval, which is used to discretize the continuous and dense combination space into an executable set of control points, so as to reduce control complexity and improve process stability.

[0050] In this embodiment, the second adhesive adhesion adjustment group is a set of candidate control parameters obtained through two-dimensional perturbation of stroke and pressure, compensation adjustment and discrete selection. Each group corresponds to a specific combination of maximum downward stroke and unit pressure, which is used for subsequent bonding experiments and scoring evaluation.

[0051] It should be noted that this embodiment incorporates mechanical structure constraints, motor output capacity, and substrate deformation limitations into a two-dimensional extension process of stroke and pressure. This expands the originally single and fixed basic adjustment parameter into a candidate adjustment set covering multiple working conditions, thereby providing a more comprehensive parameter exploration space for subsequent bonding quality evaluation and avoiding bonding instability caused by relying solely on a single empirical parameter.

[0052] It should be noted that the core of this method lies in generating parameter combinations for the boundary conditions through bidirectional perturbation, and then compensating and correcting them under physical constraints, so that all candidate parameters can be actually implemented, rather than being theoretically invalid parameters, thereby ensuring that each group of second adhesive adhesion adjustment groups has real process feasibility.

[0053] It should be noted that the beneficial effect of this technology lies in transforming the bonding process from single-point experience control to parameter space control covering multiple working conditions by structurally expanding and constraining the stroke and pressure. This provides a data basis for subsequent selection of optimal bonding parameters through a scoring mechanism, while significantly improving the adaptability to different warped substrates and different adhesive layer states.

[0054] S3, based on the gradual expansion of the contact area during the pressing process, divides the maximum pressing stroke into a contact establishment stage and a compaction stage; adjusts the motor torque output during the contact establishment stage and keeps the motor torque output stable during the compaction stage, thereby converting the adhesive adhesion adjustment group into motor torque and the end position of the pressing head.

[0055] In this embodiment, based on the gradual expansion of the contact area during the pressing process, the maximum pressing stroke is divided into a contact establishment stage and a compaction stage. The motor torque output is adjusted during the contact establishment stage, and the motor torque output is kept stable during the compaction stage. This transforms the adhesive adhesion adjustment group into motor torque and the press head termination position. Specifically: Based on the relationship between the change in pressure head displacement and the change in contact signal on the substrate surface collected during the pressing process, when the contact signal changes from a non-contact state to a continuous contact state as the pressure head displacement moves from the initial contact point to the pressing direction, the displacement position is determined as the contact start position. Based on the contact start position, the maximum downward stroke is divided into a contact establishment interval and a compaction holding interval. The contact establishment interval is the displacement range from the contact start position to the contact signal coverage area reaching the preset continuous coverage ratio. The compaction holding interval is the displacement range from the contact start position to the maximum downward stroke end position. Within the contact establishment interval, based on the continuous change of the contact signal coverage area with the increase of the pressure head displacement, the torque output of the drive motor is gradually increased to make the pressure head contact area continuously expand with the displacement. Within the compaction holding range, based on the state that the contact signal coverage area no longer changes with the addition of new areas, the torque output of the drive motor is maintained at the corresponding output value at the end of the contact establishment range, and the displacement locking control is performed on the end position of the pressure head to keep the pressure head in a constant compaction state. The gradual torque increment control process within the contact establishment interval is combined with the torque holding and termination position locking control process within the compaction holding interval, so that the adhesive adhesion adjustment group is transformed into a combined execution relationship of drive motor torque control quantity and pressure head termination position control quantity.

[0056] In this embodiment, the process of gradually expanding the contact area during the pressing process is a dynamic change process in which the pressing head gradually presses in from the initial state of not contacting the substrate and forms a contact area with the substrate surface. This process is recorded by a contact signal array or pressure feedback unit to depict the actual physical change process of the adhesive layer from initial point contact to overall pressure.

[0057] In this embodiment, the contact signal is a contact state change signal output by the pressure sensing unit at the bottom of the indenter or the capacitive contact detection unit. It is used to indicate whether effective contact has occurred between the indenter and the substrate and the change in the contact coverage area. When the signal gradually expands from discrete contact points to a continuous contact area, it is used to determine whether the contact establishment process is complete.

[0058] In this embodiment, the contact start position is the position point where the change from no contact state to continuous contact signal is first detected during the displacement of the pressure head. This position is determined by recording the displacement encoder value and the time point when the contact signal first appears continuously and stably, and is used as the reference position for the subsequent stage division.

[0059] In this embodiment, the contact establishment interval is the displacement range from the contact start position until the contact signal coverage reaches the preset continuous coverage ratio. The preset continuous coverage ratio is the proportion threshold of the effective contact area on the substrate surface reaching a continuous connection state, which is used to ensure that the adhesive layer has completed basic wetting and preliminary bonding before entering the compaction stage.

[0060] In this embodiment, the compaction holding interval is the displacement range of the pressure head from the end of the contact establishment interval to the end of the maximum pressure stroke. Within this interval, the contact area will not expand further, and the compaction of the voids inside the adhesive layer and the stabilization of the interface are achieved only by increasing the pressure.

[0061] In this embodiment, the drive motor torque output is the torque control quantity output by the motor drive system to the pressure head actuator. This control quantity is realized through motor current command and directly corresponds to the loading intensity applied by the pressure head to the substrate surface, and is used to adjust the force state of the pressure head at different contact stages.

[0062] In this embodiment, the incremental control is a control method that increases the motor torque output in stages according to the changing trend of the contact area with the expansion of displacement within the contact establishment interval. Each stage corresponds to the degree of additional coverage of the contact area, so that the pressure increases synchronously with the expansion of the contact area.

[0063] In this embodiment, the torque is maintained at a stable output level at the end of the contact establishment interval within the compaction holding interval, so that the pressure head continues to apply a constant clamping force without expanding the contact area, thereby maintaining the stability of the internal structure of the adhesive layer.

[0064] In this embodiment, the end position of the pressure head is the spatial position that the pressure head is finally locked in during the entire pressing process. This position is fixed by writing displacement holding instructions through the servo control system to ensure that the pressure head does not undergo additional displacement disturbance during the compaction stage.

[0065] In this embodiment, displacement locking control is a control method that feeds back the current position of the pressure head as the target position to the servo controller within the compaction holding range, and maintains the pressure head at that position without changing through closed-loop control, in order to prevent secondary disturbances in the adhesive layer during the compaction stage.

[0066] In this embodiment, the combined execution relationship between the drive motor torque control quantity and the pressure head termination position control quantity is to uniformly encode the dynamic torque adjustment command of the contact establishment stage and the displacement holding command of the compaction stage, so that one adhesive adhesion adjustment group simultaneously corresponds to two types of execution parameters: motor output control and mechanical displacement control, thereby realizing complete control mapping of the bonding process.

[0067] It should be noted that this embodiment introduces contact initiation detection and contact coverage ratio determination, dividing the pressing head pressing process from continuous displacement behavior into a physically meaningful contact establishment stage and a compaction holding stage, thereby changing the motor torque control from empirical setting to dynamic matching control based on the contact expansion process, thus improving the bonding consistency under different substrate warping conditions.

[0068] It should be noted that the key to this technology lies in using contact signal to drive the stage division and synchronously binding the change in contact area with the motor torque output, so that the pressure loading process no longer depends on a fixed empirical curve, but adapts to the actual contact state, thereby avoiding the problem of overpressure and false contact occurring at the same time.

[0069] It should be noted that the beneficial effect of this technology lies in transforming the backing adhesive adhesion adjustment group into a combined control relationship between motor torque and the end position of the pressure head, realizing a direct mapping from abstract parameters to executable control quantities, giving the bonding process a clear execution path, and improving the pressure head's ability to accurately control the wetting and compaction of the adhesive layer at different contact stages.

[0070] S4, convert each of the second adhesive backing adjustment groups and load them respectively to the partition pressure control unit, and perform electronic module bonding under vacuum-assisted bonding conditions.

[0071] In this embodiment, the conversion of each second adhesive adhesion adjustment group and its loading onto the partition pressure control unit, and the execution of electronic module bonding under vacuum-assisted bonding conditions, specifically involves: Each second adhesive backing adhesion adjustment group is analyzed according to the correspondence of the zone pressure control unit. The motor torque control quantity and the pressure head termination position control quantity corresponding to each adjustment group are extracted. The motor torque control quantity is converted into the drive motor current setpoint, and the pressure head termination position control quantity is converted into the servo displacement target position. The servo displacement target position is written into the motion control address of each partition of the pressure head array, and the drive motor current setpoint is written into the current control channel of each partition motor driver, so that each grid cell forms an independent displacement and driving force output channel. After the vacuum chamber reaches the set negative pressure, the control channels are synchronously triggered according to the pressure head array partition to execute the downward pressure, so that each pressure head establishes contact according to the corresponding servo displacement target position, and outputs driving force according to the corresponding motor current set value to complete the bonding during the contact process; After each pressure head reaches the target position of the servo displacement, the system switches to the hold control mode to maintain the displacement control at the termination position, maintain the motor current output at the corresponding given value, and complete the adhesive layer wetting and expansion and interface stable bonding process under the continuous action of the vacuum environment.

[0072] In this embodiment, the second adhesive adhesion adjustment group is a set of candidate control parameters generated in the previous stage. Each group is formed by perturbation, compensation and discretization of the maximum pressing stroke and unit pressure. It is used to represent the combination configuration of the pressure head movement and loading intensity under different bonding conditions, and serves as the input basis for the partition control in this step.

[0073] In this embodiment, the partitioned pressure control unit is a multi-channel control system formed by dividing the bonding area according to the pressure head array or grid unit. Each channel corresponds to an independent pressure head execution unit, which is used to receive displacement control and driving force control commands respectively, thereby realizing independent bonding control of different areas.

[0074] In this embodiment, the motor torque control quantity is a control parameter obtained in the preceding steps to represent the pressure head loading intensity. This parameter is indirectly characterized by the motor drive current. Its essence is the torque control command output by the drive motor during the bonding process, which is used to determine the pressure intensity level applied by the pressure head to the substrate.

[0075] In this embodiment, the pressure head termination position control quantity is a control parameter that represents the final spatial displacement position that the pressure head needs to maintain after completing the pressing. This parameter is obtained by converting the maximum pressing stroke of the pressure head and is used to limit the final dwell position of the pressure head in the bonding completion stage and form a stable pressing state.

[0076] In this embodiment, the drive motor current setpoint is converted into an electrical signal input value that the motor torque control quantity can be directly executed by the motor driver. This value is written through the driver current control interface to achieve direct control of the motor output torque, thereby indirectly controlling the intensity of the pressure head's action on the adhesive layer.

[0077] In this embodiment, the servo displacement target position is the target coordinate position that can be recognized by the pressure head motion control system after converting the pressure head termination position control quantity into the pressure head motion control system. This position is written into the motion control address by the servo controller, so that the pressure head can accurately stop and maintain according to the preset spatial position.

[0078] In this embodiment, the motion control address of each partition of the pressure head array is a unique address code used in the control system to identify each pressure head execution unit. Each address corresponds to a physical pressure head or pressure head group, which is used to realize multi-channel parallel displacement control and status writing.

[0079] In this embodiment, the current control channel is an independent input interface channel in the partition motor driver, which is used to receive the given value of the drive motor current and convert it into the actual output current, thereby realizing independent control of the loading intensity of each partition pressure head.

[0080] In this embodiment, the vacuum chamber is set to a negative pressure that is a pre-set pressure reduction state in the bonding environment. This state is achieved by using a vacuum pump system to reduce the internal air pressure of the chamber to a stable range lower than the external environmental pressure, thereby reducing air residue between the adhesive layer and the substrate and improving the interface contact efficiency.

[0081] In this embodiment, synchronous triggering means that after the vacuum cavity reaches the set negative pressure condition, the control system sends an execution signal to all partition pressure head control channels at the same time, so that all pressure heads start pressing down at the same time reference, thereby ensuring the synchronicity of the overall bonding process.

[0082] In this embodiment, the contact establishment process is a process in which the pressure head gradually presses down according to the target position of the servo displacement and gradually forms contact with the surface of the substrate. This process is achieved by the combined action of displacement control and current control, so that the pressure head gradually completes the transition from the idle stroke to the effective pressure state.

[0083] In this embodiment, the control mode is maintained as a stable control state after the pressure head reaches the target position of the servo displacement. In this state, the displacement control maintains the termination position unchanged, and the current control maintains the corresponding given value unchanged, so as to ensure that the adhesive layer completes the final wetting and curing stabilization process under constant pressure.

[0084] In this embodiment, the process of adhesive layer wetting and spreading and interface stabilization bonding is a process in which, under a constant vacuum environment and stable pressure, the colloid gradually fills the microscopic gaps in the substrate and eliminates residual gas, so that the contact interface gradually evolves from a local contact state to a continuous and uniform adhesion state.

[0085] It should be noted that in this embodiment, the second adhesive adhesion adjustment group is decomposed into directly executable current control quantity and displacement control quantity, and mapped to the partition control address system, so that the abstract parameters can be synchronously issued and executed in parallel in the multi-pressure head system, thereby avoiding the problem of uneven bonding caused by traditional single-channel control.

[0086] It should be noted that the key to this technology lies in the unified coordination of the vacuum environment, the partitioned pressure head array, and the dual-parameter control system, so that each pressure head has both independent displacement control capability and independent loading strength control capability, thereby achieving adaptive bonding under complex warped substrate conditions.

[0087] It should be noted that the beneficial effect of this technology is that, through partition mapping and dual-parameter execution mechanism, the second adhesive adhesion adjustment group can be directly converted into control commands that can be executed by the equipment, realizing parallel and refined control of the bonding process, while significantly improving the contact consistency and adhesive layer stability between different areas.

[0088] S5 extracts features from the bonding process data and constructs a reward signal based on the bonding results. It then generates the adhesive adhesion score for each second adhesive adhesion adjustment group through state-reward cumulative evaluation.

[0089] In this embodiment, the step of extracting features from the bonding process data and constructing a reward signal based on the bonding results, and generating an adhesive adhesion score for each second adhesive adhesion adjustment group through state-reward cumulative evaluation, specifically involves: During the bonding experiment for each second adhesive backing adhesion adjustment group, the pressure head displacement sequence, motor torque output sequence, and contact coverage state sequence were collected synchronously over time. The same adjustment group number was used as the data association identifier to bind the process data and the bonding result data one-to-one. The entire bonding process is divided into sampling windows with equal time intervals. Within each sampling window, the changes in contact coverage, displacement, and torque are statistically analyzed to form a window-level process state sequence. All windows are then arranged in chronological order to form a complete process state chain. Within each window, the change in contact coverage and the change in displacement are jointly constrained and judged. When the displacement continues to increase but the contact coverage does not expand synchronously, it is recorded as an insufficient contact establishment window event; when the contact coverage expands but the torque output continues to rise, it is recorded as an overpressure loading window event; when the contact coverage reaches stability ahead of time but the displacement continues to advance, it is recorded as an overstroke action window event. The above window events in the entire process state chain are cumulatively counted to obtain the cumulative number of insufficient contact, the cumulative number of overpressure loading, and the cumulative duration of overstroke, and a process state accumulation vector is formed. After bonding is completed, interface quality inspection is performed on the corresponding adjustment group to obtain the bubble area ratio, edge lifting length and peel strength test values, and convert them into quality return scalars according to a unified dimension. Bind the process state accumulation vector and the quality return scalar according to the same adjustment group number, so that each adjustment group forms a process state chain-quality result corresponding data pair; Based on the corresponding data pairs of all adjustment groups, the quality impact contribution value is cumulatively updated group by group according to the time window, so that each type of window event in the process state accumulation vector corresponds to a quality impact increment, and the contribution values ​​of each window are accumulated in time order to form the process return accumulation value. The cumulative value of the process return is consistent with the preset quality return scalar value to obtain the adhesive adhesion score corresponding to the second adhesive adhesion adjustment group.

[0090] In this embodiment, the second adhesive adhesion adjustment group is a set of candidate control parameters input in the bonding control stage. Each group is formed by the combination of maximum downward stroke and unit pressure, and the current control and displacement control mapping has been completed in the previous step. It is used as an index object for data association and scoring calculation in this step.

[0091] In this embodiment, the bonding experiment process is a complete execution of each second adhesive adhesion adjustment group in a vacuum partition bonding device. This process includes stages such as pressure head pressing, contact establishment, pressure holding and release, which are used to generate full-process operation data and final bonding quality results corresponding to the adjustment group.

[0092] In this embodiment, the pressure head displacement sequence is a set of pressure head position change data recorded in chronological order by a displacement sensor during the bonding process. It is used to reflect the change in the pressure depth of the pressure head at different time points and is an important basic data for judging whether the contact is established sufficiently.

[0093] In this embodiment, the motor torque output sequence is a set of data recorded by the drive system during the bonding process, showing the change of motor output intensity over time. This sequence is used to reflect the change process of pressure head loading intensity and to determine whether there is an overpressure or underloading state.

[0094] In this embodiment, the contact coverage state sequence is a recorded sequence of the effective contact area between the pressure head and the substrate over time, obtained by a contact sensor or pressure array, used to characterize the expansion process of the adhesive layer from local contact to overall bonding.

[0095] In this embodiment, the adjustment group number is an index code used to uniquely identify each second adhesive adhesion adjustment group. This number is used to bind process data, status events and final quality results one-to-one, so as to avoid data from different experimental groups from being mixed up.

[0096] In this embodiment, the sampling window is a discrete time segment formed by dividing the continuous bonding process into fixed time intervals. Each window contains information on displacement, torque and contact state changes within that time period, which is used to realize local process behavior analysis.

[0097] In this embodiment, the window-level process state sequence is a structured data set formed by statistically analyzing the displacement change, torque change, and contact coverage change within each sampling window. It is used to describe the operating state of the bonding process on a local time scale.

[0098] In this embodiment, the insufficient contact establishment window event is a situation in which the pressure head displacement continuously increases within a certain time window but the contact coverage does not expand synchronously. This event indicates that the pressure head pressing down is not effectively converted into actual contact area expansion, that is, there is a virtual stroke phenomenon.

[0099] In this embodiment, the overpressure loading window event is a window state in which the contact coverage has expanded but the motor torque continues to rise. This event indicates that the unit contact area is subjected to excessively high loading intensity, which may lead to excessive extrusion of the adhesive layer or stress concentration in the structure.

[0100] In this embodiment, the overstroke action window event is a window state in which the pressure head displacement continues to increase after the contact coverage has reached a stable state. This event indicates that the pressure head continues to press down but does not bring about effective contact improvement, which may cause excessive compression of the substrate.

[0101] In this embodiment, the process state accumulation vector is a set of structured indicators formed by performing full-process statistics on all window events. This vector consists of the cumulative number of insufficient contact, the cumulative number of overpressure loading, and the cumulative duration of overstroke action, and is used to quantify the distribution of abnormal behavior of different adjustment groups in the bonding process.

[0102] In this embodiment, interface quality detection is the process of inspecting the bonding interface using visual inspection or mechanical testing equipment after bonding is completed, in order to obtain quality result data such as air bubbles inside the adhesive layer, edge lifting, and adhesion strength.

[0103] In this embodiment, the bubble area ratio is the proportion of the area not covered by the adhesive layer in the bonding interface, which is statistically determined by image recognition and is used to reflect the degree of insufficient wetting of the adhesive layer.

[0104] In this embodiment, the edge lifting length is the cumulative length of the area where the substrate edge is not fully bonded after bonding, which is used to characterize the degree of adhesion failure in the edge area.

[0105] In this embodiment, the peel strength test value is the maximum peel force obtained by performing a peel test on the bonding interface using a tensile testing device, which is used to reflect the overall adhesion strength of the adhesive layer.

[0106] In this embodiment, the quality feedback scalar is a single value obtained by fusing the bubble area ratio, edge lifting length, and peel strength test values ​​according to a unified dimension conversion rule, and is used to uniformly represent the quality of bonding.

[0107] In this embodiment, the process state chain-quality result correspondence data pair is a data structure formed by binding the process state accumulation vector of each adjustment group with the quality reward scalar according to the number, which is used to establish the correspondence between process behavior and final quality.

[0108] In this embodiment, the quality impact increment is the impact weight value obtained by gradually accumulating and updating the quality change trend corresponding to the same window event in all adjustment group data. It is used to reflect the direction and intensity of the impact of different process events on the final bonding quality.

[0109] In this embodiment, the cumulative process return value is the result value formed by gradually accumulating the quality impact increments corresponding to each window event in chronological order, and is used to represent the comprehensive impact of the entire bonding process on the final quality.

[0110] In this embodiment, the consistency normalization process is a method of mapping the cumulative value of process rewards and the scalar value of quality rewards to a unified numerical scale, so that the two are under the same evaluation benchmark, which is used to eliminate the bias caused by the difference in the units of data from different sources.

[0111] It should be noted that this embodiment breaks down the bonding process into time window-level state events and binds each type of abnormal operation behavior to the final quality result group by group, so that the process data and result data that were originally difficult to be directly correlated form a traceable mapping relationship, thereby realizing the construction of a quantifiable evaluation basis for the bonding quality of different adjustment groups.

[0112] It should be noted that the core of this method lies in integrating three key variables—pressure head displacement, motor torque, and contact coverage—into the same state chain structure through a window-level event accumulation mechanism. This makes quality assessment no longer dependent on a single final state detection, but rather determined by the behavior throughout the entire process, thereby improving the stability and discriminative power of the evaluation.

[0113] It should be noted that the beneficial effect of this technology is that by using a state-reward accumulation method, the dynamic behavior in the bonding process is transformed into an accumulative quantifiable index, and a consistent mapping relationship is established with the final quality result. This enables a refined scoring of the second adhesive adhesion adjustment group, providing reliable data for subsequent optimal parameter selection.

[0114] S6. The maximum downward stroke and unit pressure in each second adhesive adhesion adjustment group are used as two-dimensional inputs, and the corresponding adhesive adhesion score is used as the output to fit and construct a three-dimensional evaluation surface. The optimal adhesive adhesion adjustment group is selected by searching the extreme values ​​of the surface and applied to the electronic module bonding process.

[0115] In this embodiment, the maximum downward stroke and unit pressure in each of the second adhesive adhesion adjustment groups are used as two-dimensional inputs, and the corresponding adhesive adhesion score is used as the output to fit and construct a three-dimensional evaluation surface. The optimal adhesive adhesion adjustment group is then selected through surface extremum search and applied to the electronic module bonding process. Specifically: The maximum downward stroke and unit pressure corresponding to each second adhesive adhesion adjustment group are used as two-dimensional parameter inputs, and the corresponding adhesive adhesion score is used as a scalar output to construct the adjustment group-score corresponding data set, and the data set is arranged in order by adjustment group number. For each adjustment group in the data set, extract the stroke interval identifier corresponding to its maximum downward stroke and the pressure interval identifier corresponding to the unit pressure, and map the different interval identifiers in the two-dimensional coordinate plane to form a discrete evaluation point set; Interpolation filling is performed on the scoring differences between corresponding points of adjacent adjustment groups in the discrete evaluation point set to form a continuous stroke response band in the direction of maximum downward stroke and a continuous pressure response band in the direction of unit pressure, so that the discrete scoring points are expanded into a continuous scoring distribution surface. On the continuous scoring distribution surface, the scoring change trajectories corresponding to different unit pressures under the same maximum downward stroke are sequentially connected to form a continuous change response trajectory under the fixed stroke constraint. The response trajectories corresponding to different maximum downward strokes are then superimposed layer by layer in the direction of increasing stroke to form a continuous evaluation surface structure with stroke as the first dimension of change axis, pressure as the second dimension of change axis, and score as the third dimension of response quantity. For the transition positions in the continuous evaluation surface structure where the score changes from increasing to decreasing as the stroke or pressure increases, surface gradient reverse detection processing is performed. When the score increment in the stroke direction changes from positive to negative and the score increment in the pressure direction changes from positive to negative, the corresponding surface position is marked as a set of candidate extreme value positions. For each combined parameter point in the candidate extreme value location set, a local neighborhood parameter range is constructed with it as the center in the stroke direction and pressure direction. Within this neighborhood range, the corresponding score values ​​of each neighboring point are extracted and compared for consistency. When the score of the center point is higher than the scores of all neighboring points within the neighborhood range and the fluctuation range of the score difference is in a stable range, the combined parameter point is determined as a stable extreme value point set. The set of stable extreme points is subjected to cross-neighborhood repeated verification. The number of times each extreme point maintains the highest score in multiple adjacent neighborhoods is counted. The extreme points that have appeared a preset number of consecutive confirmations are determined as the optimal adhesive adhesion adjustment group parameter pair. The optimal adhesive adhesion adjustment parameters are applied to the corresponding maximum downward stroke and unit pressure to the partition pressure control unit, and the electronic module bonding process is performed under vacuum-assisted bonding conditions.

[0116] In this embodiment, the adjustment group-score corresponding data set is a structured data set used to characterize the mapping relationship between input parameters and output performance of different second adhesive adhesion adjustment groups in the bonding experiment. The input parameters include the maximum downward stroke and unit pressure, and the output is the adhesive adhesion score. This set ensures the consistency of each group of experimental data in time and experimental batch by sorting by adjustment group number, so that the subsequent evaluation surface construction is based on the same data benchmark.

[0117] In this embodiment, the stroke interval identifier is the interval number formed after segmenting and discretizing the maximum downward stroke. This processing method divides the continuous downward stroke into several continuous intervals according to a preset resolution. The maximum downward stroke of each adjustment group falls into the corresponding interval according to its value and is assigned an interval label to eliminate the impact of single-point stroke fluctuation on the overall evaluation.

[0118] In this embodiment, the pressure range is identified by the interval number formed by dividing the unit pressure into segments according to the output capacity range of the drive motor and the process safety pressure range. This process divides the continuous pressure value into segments according to the preset pressure step range, so that the unit pressure of each adjustment group is mapped to the corresponding pressure range, thereby establishing discrete comparability in the pressure dimension.

[0119] In this embodiment, the discrete evaluation point set is a set of points formed by combining and mapping the stroke interval identifier and pressure interval identifier of each adjustment group in a two-dimensional coordinate system, where the horizontal axis is the stroke interval position and the vertical axis is the pressure interval position. Each point corresponds to an actual experimental adjustment group and carries its corresponding adhesive adhesion score.

[0120] It should be noted that the interpolation filling process for scoring differences refers to the process of extending and inferring the scores of adjacent intervals of stroke and pressure in different adjustment groups by numerically extending adjacent measured scoring points to make the scores form a continuous trend in space. This process is based on the consistency of the change direction of adjacent experimental points to extend the calculation, thereby avoiding the breakage of the evaluation surface caused by discrete experiments.

[0121] In this embodiment, the stroke response band is a scoring distribution band structure formed by continuously connecting and completing adjacent discrete scoring points in the stroke direction under fixed pressure variation conditions. It is used to reflect the continuous response trend of adhesion performance during different pressure stroke variations.

[0122] In this embodiment, the pressure response band is a distributed band structure formed by continuously connecting the scoring points corresponding to different unit pressures in the direction of increasing pressure under fixed stroke conditions. It is used to reflect the continuous influence of pressure changes on the wetting and adhesion stability of the adhesive.

[0123] In this embodiment, the continuous evaluation surface structure is a three-dimensional evaluation model formed by superimposing the stroke response band and the pressure response band. The stroke is the first variation axis, the pressure is the second variation axis, and the adhesive adhesion score is the third response axis. This surface is used to express the overall variation law of adhesion performance under different combinations of process parameters.

[0124] It should be noted that the surface gradient inverse detection process refers to comparing the local change direction of the score in the evaluation surface with the change of stroke or pressure point by point. When the score changes from an increasing state to a decreasing state, the inflection point of the change is identified and recorded as the boundary position where there may be a performance peak or performance degradation, which is used for subsequent extreme value screening.

[0125] In this embodiment, the candidate extreme value location set is a set of all surface locations that satisfy the characteristic of simultaneous score growth inflection in both the stroke direction and the pressure direction. This set is used to filter parameter combination regions that may correspond to the best adhesion effect.

[0126] In this embodiment, the local neighborhood parameter range is a parameter subspace formed by extending one or more discrete steps in the stroke direction and pressure direction, centered on the candidate extreme point, to adjacent intervals. This subspace is used to detect whether the extreme point has stable optimality in the local range.

[0127] In this embodiment, the stable extreme point set is a set of parameter points that simultaneously satisfy the condition that the score is higher than all neighboring points and the score fluctuation amplitude is less than a preset stable threshold within the local neighborhood. This set is used to eliminate pseudo extreme points caused by local noise or abnormal experiments.

[0128] In this embodiment, the cross-neighborhood repeated verification process is a process of repeatedly detecting the stability of a stable extreme point in multiple neighborhoods of different scales. By progressively expanding the neighborhood range, it is verified whether the point continues to maintain the optimal score state, thereby improving the robustness of the optimal parameter selection.

[0129] In this embodiment, the preset number of consecutive confirmations is the repeated verification threshold used to determine whether a certain extreme point is the final optimal parameter. This threshold is set by setting multiple neighboring verification windows. Only when a certain extreme point maintains the highest score in multiple windows is it confirmed as the final optimal parameter.

[0130] It should be noted that the construction of the above-mentioned three-dimensional evaluation surface is not a simple mathematical fitting, but a discrete-to-continuous transformation process based on point-by-point mapping of actual bonding experimental data. By introducing interval mapping, neighborhood consistency verification and cross-scale verification mechanisms, the evaluation results can simultaneously reflect the comprehensive influence of process parameters on the sufficiency of contact establishment, colloidal flow stability and interface residual stress distribution, thereby avoiding parameter misjudgment caused by a single experimental point.

[0131] It should be noted that this method combines the maximum downward stroke with the unit pressure and introduces a stable extreme value screening mechanism, enabling the system to extract the optimal combination of process parameters with global consistency from discrete experimental data, thereby improving the ability to suppress bubbles, control edge warping, and maintain peel strength consistency during the bonding process of electronic modules.

[0132] It should be noted that the beneficial effect of this technology lies in constructing a continuous evaluation surface driven by experimental data, and combining neighborhood consistency and cross-scale verification screening mechanisms, so that the selection of bonding parameters is transformed from an empirical rule-based approach to a verifiable optimal solution search approach, thereby significantly improving the adhesion stability and process transferability of the adhesive under different order conditions.

[0133] Example 2: This invention also includes a data-driven adaptive adjustment system for adhesive adhesion parameters, comprising an adhesive adhesion initialization module, an adjustment disturbance generation module, a torque mapping module, a partitioned execution module, a feedback scoring module, and a surface optimization module. The adhesive adhesion initialization module acquires hardware constraint data and order demand data, constructs a stroke-pressure feasible region, and generates a combination of the basic maximum downward stroke and the basic unit pressure through constraint optimization to construct a basic adhesive adhesion adjustment group. The adjustment disturbance generation module continuously perturbs the basic adhesive adhesion adjustment group bidirectionally in both the stroke and pressure dimensions, generating several second adhesive adhesion adjustment groups covering different contact adequacies and adhesive flow states. The torque mapping module, based on the gradual expansion of the contact area during the pressure head's downward movement, divides the maximum downward stroke into a contact establishment stage and a compaction stage. The process involves several stages: During the contact establishment stage, the motor torque output is adjusted; during the compaction stage, the motor torque output is kept stable, thus converting the adhesive adhesion adjustment group into motor torque and pressure head termination position; a zone execution module converts each second adhesive adhesion adjustment group and loads it to the zone pressure control unit, executing electronic module bonding under vacuum-assisted bonding conditions; a feedback scoring module extracts features from the bonding process data and constructs a feedback signal based on the bonding results, generating an adhesive adhesion score for each second adhesive adhesion adjustment group through state-feedback cumulative evaluation; and a surface optimization module uses the maximum downward stroke and unit pressure of each second adhesive adhesion adjustment group as two-dimensional inputs, and the corresponding adhesive adhesion score as the output to fit and construct a three-dimensional evaluation surface, then uses surface extremum search to select the optimal adhesive adhesion adjustment group for application in the electronic module bonding process.

[0134] The present invention also includes an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform a data-driven adaptive adjustment method for adhesive backing parameters.

[0135] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements a data-driven adaptive adjustment method for adhesive backing parameters.

[0136] In the embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0137] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0138] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0139] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0140] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0142] In the embodiments provided in this disclosure, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0143] It should be noted that, in this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0144] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A data-driven adaptive adjustment method for adhesive adhesion parameters, characterized in that, Includes the following steps: Obtain hardware limitation data and order demand data, construct the stroke-pressure feasible region, and generate a combination pair of basic maximum downward stroke and basic unit pressure through constraint optimization to construct a basic adhesive adhesion adjustment group, specifically: The hardware limitation data includes the upper limit of the stroke of the pressure head structure of the electronic module bonding equipment and the rated torque of the drive motor; the order requirement data includes the substrate thickness and warpage statistical range in the order. The maximum allowable pressing depth set is determined based on the upper limit of the pressure head structure stroke, and the maximum output contact pressure set is determined based on the rated torque of the drive motor. The maximum allowable pressing depth set and the maximum output contact pressure set are then combined using a Cartesian combination to generate the initial stroke-pressure combination space. The initial stroke-pressure combination space is screened based on the substrate thickness and warpage statistical range. Stroke combinations that are less than the maximum warpage height of the substrate and correspond to insufficient contact are eliminated, and pressure combinations that exceed the allowable range of substrate compression deformation and correspond to overpressure risk are eliminated, thus obtaining a set of feasible combinations that meet the physical contact conditions. In the feasible combination set, the unit pressure and the contact coverage of the pressure head are used as constraints. The contact coverage is set to be no less than the effective bonding ratio of the substrate surface. Combinations with unit pressure exceeding the critical range of substrate compression are eliminated, while the combination range that meets the coverage and compression limits is retained to form the stroke-pressure feasible region. Within the feasible range of stroke-pressure, the difference between the pressure head termination displacement and the remaining thickness of the substrate after compression is used as the screening criterion. Combination points whose difference falls within the preset allowable range are preferentially retained. The combination points are then discretely selected to generate combination pairs of the maximum downward stroke and the unit pressure of the base, thus constructing the base adhesive adhesion adjustment group. For the basic adhesive adhesion adjustment group, bidirectional continuous perturbation is performed in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacy and adhesive flow states. Based on the gradual expansion of the contact area during the pressing process, the maximum pressing stroke is divided into a contact establishment stage and a compaction stage. The motor torque output is adjusted during the contact establishment stage, and the motor torque output is kept stable during the compaction stage, thereby converting the adhesive adhesion adjustment group into motor torque and the end position of the pressing head. Each of the second adhesive backing adjustment groups is converted and loaded into the partition pressure control unit respectively, and electronic module bonding is performed under vacuum-assisted bonding conditions; Feature extraction is performed on the bonding process data, and a reward signal is constructed based on the bonding results. The adhesive adhesion score corresponding to each second adhesive adhesion adjustment group is generated through state-reward cumulative evaluation. The maximum downward stroke and unit pressure in each second adhesive adhesion adjustment group are used as two-dimensional inputs, and the corresponding adhesive adhesion score is used as the output to fit and construct a three-dimensional evaluation surface. The optimal adhesive adhesion adjustment group is selected by searching the extreme values ​​of the surface and applied to the electronic module bonding process.

2. The data-driven adaptive adjustment method for adhesive adhesion parameters according to claim 1, characterized in that, The basic adhesive adhesion adjustment group is subjected to bidirectional continuous perturbation in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacies and adhesive flow states, specifically: The bidirectional continuous disturbance extends the displacement along the allowable range of the mechanical clearance of the pressure head, and the pressure extends the load along the range of the output capacity of the drive motor. For the maximum downward stroke in the basic adhesive adhesion adjustment group, the displacement is extended in the increasing and decreasing directions according to the preset step size along the allowable range of the mechanical clearance of the pressure head to obtain the stroke disturbance sequence; For the unit pressure in the basic adhesive adhesion adjustment group, the load is expanded in the increasing and decreasing directions according to the preset step size along the output capacity range of the drive motor to obtain the pressure disturbance sequence. The stroke disturbance sequence and the pressure disturbance sequence are paired one by one to form a stroke-pressure disturbance combination set. The contact compensation adjustment is performed on the indenter termination displacement corresponding to each combination so that the termination displacement is not less than the maximum warping height of the substrate and does not exceed the upper limit of the indenter structure stroke. At the same time, the unit pressure corresponding to each combination is adjusted to ensure that the unit pressure does not exceed the pressure limit corresponding to the compressibility range of the substrate. After completing the displacement compensation adjustment and pressure loading limit adjustment, each combination is discretely selected at preset intervals to obtain several combination points covering different pressing depths and loading pressure ranges. Then, using the maximum pressing stroke and unit pressure corresponding to each combination point as parameter pairs, several second adhesive adhesion adjustment groups are constructed.

3. The data-driven adaptive adjustment method for adhesive adhesion parameters according to claim 2, characterized in that, Based on the gradual expansion of the contact area during the pressing process, the maximum pressing stroke is divided into a contact establishment stage and a compaction stage. The motor torque output is adjusted during the contact establishment stage, and the motor torque output is kept stable during the compaction stage. This transforms the adhesive adhesion adjustment group into motor torque and the press head termination position, specifically: Based on the relationship between the change in pressure head displacement and the change in contact signal on the substrate surface collected during the pressing process, when the contact signal changes from a non-contact state to a continuous contact state as the pressure head displacement moves from the initial contact point to the pressing direction, the displacement position is determined as the contact start position. Based on the contact start position, the maximum downward stroke is divided into a contact establishment interval and a compaction holding interval. The contact establishment interval is the displacement range from the contact start position to the contact signal coverage area reaching the preset continuous coverage ratio. The compaction holding interval is the displacement range from the contact start position to the maximum downward stroke end position. Within the contact establishment interval, based on the continuous change of the contact signal coverage area with the increase of the pressure head displacement, the torque output of the drive motor is gradually increased to make the pressure head contact area continuously expand with the displacement. Within the compaction holding range, based on the state that the contact signal coverage area no longer changes with the addition of new areas, the torque output of the drive motor is maintained at the corresponding output value at the end of the contact establishment range, and the displacement locking control is performed on the end position of the pressure head to keep the pressure head in a constant compaction state. The gradual torque increment control process within the contact establishment interval is combined with the torque holding and termination position locking control process within the compaction holding interval, so that the adhesive adhesion adjustment group is transformed into a combined execution relationship of drive motor torque control quantity and pressure head termination position control quantity.

4. The data-driven adaptive adjustment method for adhesive adhesion parameters according to claim 3, characterized in that, The process of converting each of the second adhesive adhesion adjustment groups and loading them separately into the zoned pressure control unit, and performing electronic module bonding under vacuum-assisted bonding conditions, specifically involves: Each second adhesive backing adhesion adjustment group is analyzed according to the correspondence of the zone pressure control unit. The motor torque control quantity and the pressure head termination position control quantity corresponding to each adjustment group are extracted. The motor torque control quantity is converted into the drive motor current setpoint, and the pressure head termination position control quantity is converted into the servo displacement target position. The servo displacement target position is written into the motion control address of each partition of the pressure head array, and the drive motor current setpoint is written into the current control channel of each partition motor driver, so that each grid cell forms an independent displacement and driving force output channel. After the vacuum chamber reaches the set negative pressure, the control channels are synchronously triggered according to the pressure head array partition to execute the downward pressure, so that each pressure head establishes contact according to the corresponding servo displacement target position, and outputs driving force according to the corresponding motor current set value to complete the bonding during the contact process; After each pressure head reaches the target position of the servo displacement, the system switches to the hold control mode to maintain the displacement control at the termination position, maintain the motor current output at the corresponding given value, and complete the adhesive layer wetting and expansion and interface stable bonding process under the continuous action of the vacuum environment.

5. The data-driven adaptive adjustment method for adhesive adhesion parameters according to claim 4, characterized in that, The process involves feature extraction from the bonding process data and the construction of a reward signal based on the bonding results. A state-reward cumulative evaluation is then used to generate an adhesive adhesion score for each second adhesive adhesion adjustment group. Specifically: During the bonding experiment for each second adhesive backing adhesion adjustment group, the pressure head displacement sequence, motor torque output sequence, and contact coverage state sequence were collected synchronously over time. The same adjustment group number was used as the data association identifier to bind the process data and the bonding result data one-to-one. The entire bonding process is divided into sampling windows with equal time intervals. Within each sampling window, the changes in contact coverage, displacement, and torque are statistically analyzed to form a window-level process state sequence. All windows are then arranged in chronological order to form a complete process state chain. Within each window, the change in contact coverage and the change in displacement are jointly constrained and judged. When the displacement continues to increase but the contact coverage does not expand synchronously, it is recorded as an insufficient contact establishment window event; when the contact coverage expands but the torque output continues to rise, it is recorded as an overpressure loading window event; when the contact coverage reaches stability ahead of time but the displacement continues to advance, it is recorded as an overstroke action window event. The above window events in the entire process state chain are cumulatively counted to obtain the cumulative number of insufficient contact, the cumulative number of overpressure loading, and the cumulative duration of overstroke, and a process state accumulation vector is formed. After bonding is completed, interface quality inspection is performed on the corresponding adjustment group to obtain the bubble area ratio, edge lifting length and peel strength test values, and convert them into quality return scalars according to a unified dimension. Bind the process state accumulation vector and the quality return scalar according to the same adjustment group number, so that each adjustment group forms a process state chain-quality result corresponding data pair; Based on the corresponding data pairs of all adjustment groups, the quality impact contribution value is cumulatively updated group by group according to the time window, so that each type of window event in the process state accumulation vector corresponds to a quality impact increment, and the contribution values ​​of each window are accumulated in time order to form the process return accumulation value. The cumulative value of the process return is consistent with the preset quality return scalar value to obtain the adhesive adhesion score corresponding to the second adhesive adhesion adjustment group.

6. The data-driven adaptive adjustment method for adhesive adhesion parameters according to claim 5, characterized in that, The process involves using the maximum downward stroke and unit pressure in each of the second adhesive adhesion adjustment groups as two-dimensional inputs, and the corresponding adhesive adhesion score as the output to fit and construct a three-dimensional evaluation surface. The optimal adhesive adhesion adjustment group is then selected through surface extremum search and applied to the electronic module bonding process. Specifically: The maximum downward stroke and unit pressure corresponding to each second adhesive adhesion adjustment group are used as two-dimensional parameter inputs, and the corresponding adhesive adhesion score is used as a scalar output to construct the adjustment group-score corresponding data set, and the data set is arranged in order by adjustment group number. For each adjustment group in the data set, extract the stroke interval identifier corresponding to its maximum downward stroke and the pressure interval identifier corresponding to the unit pressure, and map the different interval identifiers in the two-dimensional coordinate plane to form a discrete evaluation point set; Interpolation filling is performed on the scoring differences between corresponding points of adjacent adjustment groups in the discrete evaluation point set to form a continuous stroke response band in the direction of maximum downward stroke and a continuous pressure response band in the direction of unit pressure, so that the discrete scoring points are expanded into a continuous scoring distribution surface. On the continuous scoring distribution surface, the scoring change trajectories corresponding to different unit pressures under the same maximum downward stroke are sequentially connected to form a continuous change response trajectory under the fixed stroke constraint. The response trajectories corresponding to different maximum downward strokes are then superimposed layer by layer in the direction of increasing stroke to form a continuous evaluation surface structure with stroke as the first dimension of change axis, pressure as the second dimension of change axis, and score as the third dimension of response quantity. For the transition positions in the continuous evaluation surface structure where the score changes from increasing to decreasing as the stroke or pressure increases, surface gradient reverse detection processing is performed. When the score increment in the stroke direction changes from positive to negative and the score increment in the pressure direction changes from positive to negative, the corresponding surface position is marked as a set of candidate extreme value positions. For each combined parameter point in the candidate extreme value location set, a local neighborhood parameter range is constructed with it as the center in the stroke direction and pressure direction. Within this neighborhood range, the corresponding score values ​​of each neighboring point are extracted and compared for consistency. When the score of the center point is higher than the scores of all neighboring points within the neighborhood range and the fluctuation range of the score difference is in a stable range, the combined parameter point is determined as a stable extreme value point set. The set of stable extreme points is subjected to cross-neighborhood repeated verification. The number of times each extreme point maintains the highest score in multiple adjacent neighborhoods is counted. The extreme points that have appeared a preset number of consecutive confirmations are determined as the optimal adhesive adhesion adjustment group parameter pair. The optimal adhesive adhesion adjustment parameters are applied to the corresponding maximum downward stroke and unit pressure to the partition pressure control unit, and the electronic module bonding process is performed under vacuum-assisted bonding conditions.

7. A system using the data-driven adaptive adjustment method for adhesive adhesion parameters as described in any one of claims 1-6, characterized in that, It includes an adhesive adhesion initialization module, a disturbance generation module, a torque mapping module, a partition execution module, a feedback scoring module, and a surface optimization module; The adhesive adhesion initialization module is used to acquire hardware limitation data and order demand data, construct the stroke-pressure feasible region, and generate a combination pair of basic maximum downward stroke and basic unit pressure through constraint optimization to construct the basic adhesive adhesion adjustment group. The adjustment disturbance generation module is used to continuously disturb the basic adhesive adhesion adjustment group in both the stroke and pressure dimensions to generate several second adhesive adhesion adjustment groups covering different contact adequacy and adhesive flow states. The torque mapping module is used to divide the maximum pressing stroke into a contact establishment stage and a compaction stage based on the gradual expansion of the contact area during the pressing process. The motor torque output is adjusted during the contact establishment stage and the motor torque output is kept stable during the compaction stage, thereby converting the adhesive adhesion adjustment group into motor torque and the end position of the pressing head. The partition execution module is used to convert each of the second adhesive adhesion adjustment groups and load them to the partition pressure control unit respectively, so as to perform electronic module bonding under vacuum-assisted bonding conditions; The feedback scoring module is used to extract features from the bonding process data and construct a reward signal based on the bonding results. It generates the adhesive adhesion score corresponding to each second adhesive adhesion adjustment group through state-reward cumulative evaluation. The surface optimization module is used to fit and construct a three-dimensional evaluation surface by taking the maximum downward stroke and unit pressure in each second adhesive adhesion adjustment group as two-dimensional inputs and the corresponding adhesive adhesion score as outputs. The optimal adhesive adhesion adjustment group is then selected through surface extreme value search and applied to the electronic module bonding process.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the data-driven adaptive adjustment method for adhesive backing parameters as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the data-driven adaptive adjustment method for adhesive backing parameters as described in any one of claims 1 to 6.

Images